Great strides have been made in the interpretation of covalent histone modifications regarding their role in transcriptional regulation. Histone lysine methylation has been found to affect the structure of chromatin thereby establishing complex patterns of gene expression . In some cases, these patterns are clearly defined. For example, H3K4 methylation is most often associated with the establishment of euchromatin and the consequent activation of local gene expression . Reciprocally, methylation at H3K9 is commonly involved with the formation of heterochromatin and the ensuing silencing of nearby gene transcription [3, 4].
Initial data on the yeast HKMT, Set2, indicated that it functions in transcriptional repression by methylating H3K36 . However, the HKMT activity of Set2 was later linked to the elongation phase of RNA polymerase II (RNAPII) [25, 26]. Likewise, in a more contemporary study, an analysis of the distribution of H3K36 methylation in metazoans correlated this modification with actively transcribed genes . Most recently, methylation of H3K36 by Set2 has been associated with the recruitment of a histone deacetylase complex, Rpd3 . The overall role and implications of histone deacetylation within the coding regions of active genes is still unknown.
In mammalian epigenetics, NSD1 was one of the first HKMTs reported to act on H3K36 . Whether NSD1 functions in the activation or repression of transcription has yet to be determined. A recent investigation reported that the human HYPB protein methylates H3K36 and that this enzymatic activity is required for the role of HYPB as a transcriptional activator .
Our findings introduce Smyd2 as an H3K36-specific HKMT that acts as a transcriptional repressor. Clearly, there are other transcriptional regulatory mechanisms at work in conjunction with the methylation of H3K36. It seems that the more we learn about where histone marks are localized and what proteins facilitate the process, the less we are certain about how such localization ultimately contributes to gene regulation. Although this complicates our ability to apply a broad interpretation of histone modifications, it provides a clear direction for the pursuit of a deeper fold in the "histone code."
Smyd2 regulatory functions
Transcriptional assays demonstrated that Smyd2 can repress transcription from a luciferase reporter gene (Fig. 3B). A recent study in yeast demonstrated that methylation of H3K36 recruits a histone deacetylase complex, Rpd3 . Concurrently, another group concluded that H3K36 methylation-induced recruitment of an Rpd3 complex resulted in the reversal of lysine acetylation related to the elongation phase of RNAPII, suggesting that it functioned to stem intragenic transcription initiation . This is reminiscent of the mechanism by which the FACT complex functions. That is, as the elongation complex traverses a coding region, FACT facilitates both destabilization of the chromatin structure, to impart efficient and processive elongation, as well as reorganization of the chromatin to prevent intragenic initiation of transcription . Whereas H3K36 methylation recruits the Rpd3 complex, it has been suggested that FACT recruitment may occur through its association with CHD1, which recognizes trimethylated H3K4 . As the Rpd3 complex is known to contain Sin3 , it was particularly informative to find that Smyd2 also associates with Sin3. It will be of further interest to determine whether in vivo recruitment of Sin3 requires H3K36 methylation, the presence of Smyd2, or both.
Over-expression of Smyd2 in NIH3T3 cells significantly reduces cell growth. In a previous study, cell proliferation assays demonstrated that Smyd3 augmented cell growth when introduced into NIH3T3 cells . It is well established that cell proliferation and differentiation are coordinated by synchronized patterns of gene transcription. In the case of Smyd3, enhancement of cell growth has been shown to be dependent upon the H3K4-specific HKMT activity of the Smyd3 protein . It will be informative to determine whether the suppressive effect of Smyd2 on cell growth requires its function as an H3K36-specific HKMT. Such a determination, in tandem with identification of putative gene target specificity of Smyd2 will provide a broader mechanistic model of how the Smyd family may function.
Histone lysine methylation is more stable than other known post-translational modifications, persisting as long as several rounds of cell division [31–33]. This makes lysine methylation potentially valuable in diverse, long-lasting signaling networks, not only in the nucleus for histone and non-histone proteins, such as p53 and TAF10, but conceivably in the cytoplasm. The findings that Smyd1 and Smyd3 can localize in the cytoplasm [21, 8] along with our observation that Smyd2 is also capable of cytosolic localization, lends credence to this idea. This argument is further strengthened by the finding that Smyd1 moves from the nucleus to the cytoplasm during myogenic differentiation . Another SET domain-containing HKMT, Ezh2 and its partners Eed and Suz12, reside primarily in the cytoplasm of various mouse and human cells [34, 35]. Within the nucleus, the Ezh2 complex catalyzes H3K27 methylation, whereas the cytosolic Ezh2 binds Vav1, a controller of Rho family GTPases, and Ezh2 is important for signaling events previously attributed to Vav1 [34–36]. There is no evidence that Ezh2 methylates Vav1, so the significance of lysine methylation in the cytoplasm remains unclear. However, we are currently testing the role of Smyd2-mediated lysine methylation in the formation of stable and potentially heritable cytosolic signaling complexes with Smyd2 interaction partners and we plan to track these complexes, once formed, within resting and dividing cells.
The Smyd family
The Smyd HKMTs are set apart from other such chromatin modifying enzymes by the split nature of their SET domains. The SET domain of each Smyd protein is divided by a MYND domain (Figure 1A &1C), a zinc-finger motif that mediates protein-protein interactions. This domain is found in several transcriptional regulators shown to mediate distinct biological functions [37, 38]. For example, the MYND domain of Smyd1 is essential for its interaction with the muscle-specific transcription factor, skNAC . Additionally, ETO, a common target of chromosomal translocations in acute myeloid leukemia, directs transcriptional repression through an intact MYND domain . Thus, the importance of the MYND domain in gene regulation has been well established and it may provide some insight into other mechanisms at work through Smyd2 that affect the overall outcome of its activity in transcriptional regulation. The complete function of Smyd2 in vivo is likely dependent upon other proteins and complexes, in addition to HDAC1 and Sin3A, with which it associates. We are currently screening several other candidate interaction partners whose nature may give further clues to mechanisms and pathways regulated by Smyd2.
Northern blot analysis revealed that Smyd2 and Smyd3 are expressed in a wide variety of tissues (Fig. 1B) whereas Smyd1 is more restricted in its tissue distribution . Studies of ours and others on Smyds1-3 suggest that Smyd family members function through a common mechanism, specifically, lysine methylation. It is reasonable to assume that individual Smyd proteins associate with different transcription factors and other effector proteins that ultimately dictate specific gene regulation. No other Smyd family member is functionally redundant with Smyd1, since homozygous Smyd1-null mice are embryonic lethal at day E10.0 as a result of impaired cardiomyocyte differentiation . The significant expression of Smyd2 in the in embryonic heart (Fig. 1B, E) suggests that as with Smyd1, Smyd2 may regulate cardiac development. The identification of the biological functions of Smyd family proteins will undoubtedly reveal new insights into the relationships between chromatin modifications and the development and differentiation of specific tissues.